Effects of Hemostasis Correction on Vascular Remodeling and Synovial Gene Expression Changes in Mice after Induced Hemarthrosis

Introduction

Repeated joint bleeding in patients with hemophilia leads to hemophilic arthropathy (HA), which cannot be entirely prevented by clotting factor replacement. Vascular remodeling and permeability are associated with hemarthrosis and may contribute to HA progression; however, the mechanisms and effects of hemostasis correction are poorly understood. Here, we explored synovial vascular and gene expression changes in FVIII-deficient mice after induced hemarthrosis +/- FVIII replacement, and in wild-type mice with reversible hemostasis suppression.

Methods

Hemarthrosis was induced in FVIII-deficient mice by sub-patellar needle puncture +/- 100-200 IU/kg recombinant human FVIII (rhFVIII) 2 hours before and 6 hours after injury. Wild-type mice were treated with 10 µg/ml warfarin for 7 days and 0.25 mg/kg anti-FVIII 2 hours before injury (^hypoBALB/c). After injury, warfarin was continued for 2 weeks or reversed on day 2 with 100 IU/kg 4-factor prothrombinase complex concentrate (4F-PCC). Vascularity and gene expression were analyzed at baseline and 2 weeks post-injury. Vessel number and remodeling were assessed by histology with Safranin-O-Fast Green and α-smooth muscle actin (αSMA) staining, respectively, and microvascular flow was detected by musculoskeletal ultrasound with Power Doppler. The permeability of synovial vessels was determined by quantification of extravasated albumin-bound Evans blue dye in knee joints. For synovial gene expression studies, libraries were prepared using the NEBNext Ultra II DNA Library Prep Kit and sequenced on an Illumina NextSeq500 (single-end; 75bp reads). The limma-voom method (R BioConductor) was used for differential expression analyses and functional enrichment was performed using Signaling Pathway Impact Analysis.

Results

In FVIII-deficient mice, knee injury caused profound hemarthrosis that was largely prevented with rhFVIII prophylaxis (day 2 hematocrit: 26.4% and 46.3%). Soft tissue proliferation increased to a similar extent, as did various vascular parameters: microvascular flow (vehicle: 1.8-fold; rhFVIII: 1.5-fold), vessel number (vehicle: 2.3-fold, p=0.0005; rhFVIII: 2.0-fold, p=0.004), vessels with diameter ≥ 20 µm (vehicle: 2.9-fold, p=0.02; rhFVIII: 2.7-fold, p=0.02), and αSMA area per vessel (vehicle: 2.3-fold, p>0.05; rhFVIII: 3.6-fold, p=0.0006). Vascular permeability also increased significantly (1.7-fold, p=0.0007) and was only partially rescued by rhFVIII prophylaxis (1.3-fold, p>0.05). In ^hypoBALB/c mice (day 2 hematocrit: 29.6%), significant but less pronounced vascular changes occurred regardless of hemostasis correction, and without associated permeability, suggesting this is uniquely associated with FVIII-deficiency. RNA sequencing in FVIII-deficient mice revealed a strong transcriptional response to hemarthrosis (1527 differentially expressed genes (DEG), 13 perturbed pathways) that was only partially dampened with rhFVIII treatment (891 DEG, 20 pathways). Perturbation of extracellular matrix (ECM)-receptor interactions was highly significant in both groups (vehicle: pNDE=7.7x10^-10; rhFVIII: pNDE=6.8x10^-9). Similarly, numerous genes relating to angiogenesis and ECM remodeling, including collagens and MMPs, were up-regulated and minimally affected by rhFVIII treatment. These transcriptional changes may facilitate the observed vascular remodeling after hemarthrosis.

Conclusions

Hemarthrosis triggers profound changes in synovial gene expression, notably ECM components, that may drive the associated soft tissue and vascular changes, including vessel remodeling and leakiness. These processes are incompletely mitigated by hemostasis correction and may exacerbate (re-) bleeding tendencies. Therefore, further exploration is needed to identify key molecular pathways that can be targeted to intercept the progression of HA.